Development and evaluation of empirical and dynamic models for predicting passage rate in growing Saanen goats

• Grant number: 14/09229-3
• Support Opportunities: Scholarships abroad - Research Internship - Master's degree
• Start date: July 01, 2014
• End date: November 30, 2014
• Field of knowledge: Agronomical Sciences - Animal Husbandry - Animal Nutrition and Feeding
• Principal Investigator: Izabelle Auxiliadora Molina de Almeida Teixeira
• Grantee: Thiago Francisco Ventoso Bompadre
• Supervisor: Normand R. St-Pierre
• Host Institution: Faculdade de Ciências Agrárias e Veterinárias (FCAV). Universidade Estadual Paulista (UNESP). Campus de Jaboticabal. Jaboticabal , SP, Brazil
• Institution abroad: Ohio State University, Columbus, United States
• Associated to the scholarship: 13/06845-2 - Development and evaluation of mathematical models for predicting the rate of passage in growing goats, BP.MS

Abstract

Passage rates has been associated to the maximum dry matter intake, feed digestibility, efficiency of ruminal microbiota, and other factors, impacting on the amount of nutrients reaching the intestine for absorption and consequently in animal production. The literature presents limited and not conclusive knowledge on the kinetics of particles passage in goats and its changes throughout the development of the animal. Some studies have been studying ways to modelling the passage rate to apply in feeding systems, in this sense some models to predict ruminants passage rates have been developed, but these passage rates models derived from other species and when used for goats becomes a erroneous practice. Thus, further studies to develop mathematical models that adequately estimate the passage rates in goats are necessary. Therefore, this study aimed to develop and evaluate empiric and dynamic models to estimate the passage rates in growing Saanen goats. The database used to develop the passage rates models is from the information of FAPESP`s project number 2008/57302-0. In this project 3 trials were conducted (5 to 15 kg of body weight , representing the weaning and post- weaning period; 15 to 30 kg body weight and 30 to 45 kg body weight, representing the final growth stage), which included 254 animals. The models will be developed using PROC MIXED (SAS, 2002), which will be tested matrices of variation and co-variation, looking for one that provides the best fit of the model. Dynamic models will be developed based on the concept that the animal can control the flow through the rumen according to his nutrients needs and the diet characteristics. The sensitivity analysis will be performed using the Monte Carlo technique. The models will be evaluate with an independent database and the accuracy of the model will be evaluate with the determination coefficient (R2) of the linear regression between predicted and observed values, the mean square error prediction (MSEP) and the concordance correlation coefficient (CCC), and its separation in mean and linear bias. We will also test the models proposed by Cannas & Van Soest (2000) and Tedeschi et al. (2011). (AU)